Bats Detectors

Acta Chiropterologica, 3(2): xxx–xxx, 2001
PL ISSN 1508-1109 © Museum and Institute of Zoology PAS










SHORT NOTES

Choosing the ‘correct’ bat detector — a reply

CHRIS CORBEN1 and GARY M. FELLERS2

1P.O. Box 2323, Rohnert Park, California CA 94927 USA; E-mail: corben@hoarybat.com
2Western EcologicalResearch Center, USGS, Point Reyes National Seashore, Point Reyes, CA 94956 USA

Key words: bat, echolocation, zero-crossing, Anabat, harmonics


Fenton (2000) compared two bat call analysis systems, one based on time-expan- sion and spectral analysis (Pettersson D980 with BatSound Pro software, www.bahn- hof.se/~pettersson/) and the other based on zero crossings analysis (ZCA — Anabat IIdetector with ZCAIM and Anabat6 soft- ware, www.titley.com.au/index.html). This author concluded that the Anabat system was significantly inferior and claimed it was insensitive, inaccurate, did not detect all species, and was not reliable for “species that vary the harmonic content of their echolocation calls.” He claimed that using Anabat resulted in “substantial sacrifice in the areas ofsensitivity and accuracy of data obtained about echolocation calls.” We contend that there are serious flaws in Fenton’s approach, and that the value judg- ments he has made in comparing two very different systems profoundly misrepresents the real value of Anabat for the purposes for which most people would want to use it.


Sensitivity

Time-expansion and spectral analysis is inherentlymore sensitive than ZCA, just as
a narrow band system using heterodyne down-conversion will be more sensitive than either. If sensitivity was the only issue, a heterodyne system should be preferred.
Greater sensitivity will allow the user to sample a larger volume of space with a sin- gle detector. In situations where bats are rare and pass by one at a time, this might be an advantage.However, the Anabat system has proven sufficiently sensitive for most purposes and, in practice, there often is not time to fully process even the Anabat data
— a situation which would be greatly aggravated by the use of a time-expansion system.
In the case of time-expansion, the greater sensitivity is more than counterbal- anced by the fact that it is possible to record bats only 7.5% of the time,assuming the detector is being constantly watched by an experienced observer (12 s of data every
160 s — Fenton, 2000). Even if we accept Fenton’s assertion that the time-expansion system is five times more sensitive than the ZCA, the latter would still detect 2.7 times as many bats in such a setting because it operates almost continuously (100% duty cycle).
2 Short Notes

We see no basisfor Fenton’s claim that differences in the sensitivity of the two sys- tems will “further complicate the matter of natural variability” or that variability in the calls produced by bats can somehow com- bine with variations in detector sensitivity to “affect the quality of data.” A very basic skill in effective use of any bat detector sys- tem is the ability to make the distinction betweenvariation due to the bat and varia- tion due to how well a call is received by the detector. The latter is not only affected by the sensitivity of the detector, but also by the distance between detector and bat, which is nearly always changing. No matter how sensitive a detector is, it will still detect many calls very poorly, since there will always be bats which are too far away to detectproperly. The total range of varia- tion in reception should not be expected to depend on detector sensitivity.
Although we would not be surprised, or disturbed, to find that Anabat detects only one fifth as many calls as the time-expan- sion system, we have no confidence in Fenton’s quantification of this difference. Fenton’s paper implies the presence of a clear dichotomy between calls detected...